Method and system for predicting tone of Lhasa dialect of Tibetan language

Abstract

The invention discloses a method and a system for predicting tone of Lhasa dialect of Tibetan language. The method includes acquiring continuous Tibetan language texts; segmenting the texts to obtain segmented word strings; determining syllable tone prediction characteristics of various syllables in segmented words; determining tonal patterns of the syllables according to the syllable tone prediction characteristics of the syllables by the aid of syllable tonal pattern prediction models obtained by means of training in advance. Each segmented word string comprises the single corresponding segmented word or the multiple corresponding segmented words. The method and the system have the advantage that the syllable tone prediction accuracy can be improved for the Lhasa dialect in the continuous Tibetan language texts.

Description

technical field [0001] The present invention relates to the technical field of text processing, in particular to a method and system for predicting Tibetan Lhasa tone. Background technique [0002] In recent years, with the development of the needs of the information society, users have put forward higher requirements for human-computer interaction, and high-natural speech synthesis effects have become an important symbol of high-performance speech synthesis systems. [0003] Speech synthesis is also called text-to-speech (Text to Speech) technology. The text-to-speech conversion process is to convert text sequences into speech unit sequences first, and then the system generates speech waveforms according to the speech unit sequences. In the traditional speech synthesis system, text regularization, word segmentation, prosodic phrase division, etc. are mainly realized at the front end of the synthesis, and then the appropriate speech splicing unit or synthesis parameters are ...

AI Technical Summary

Problems solved by technology

[0008] 3. Rules-based monosyllabic tone determination and bisyllable continuous tone sandhi determination often have a large number of counterexamples in continuous Tibetan Lhasa

[0009] Obviously, the traditional rules-based definition of Tibetan Lhasa tone tone and syllable-linked tone-modification methods are only suitable for a small amount of data, and cannot meet the needs of large-vocabulary continuous speech synthesis.

Images